4-4 FOREST MONITORING
Forest ecosystems represent important economic and environmental assets which are to be monitored along a range of appropriate scales. In many parts of the world, forests offer an appreciable source of income from timber and non-wood products. Woody ecosystems are also large reservoirs of carbon which can be drastically manipulated with consequences for the carbon cycle and atmospheric chemistry. The forest cover also offers a certain level of protection over soil and water resources, and those ecosystems are home to a very large variety of life-forms. In many countries, the forest land is often seen as the last frontier for the expansion of agriculture and as such is subject of often rapid and radical transformations. Given these many perspectives on forests and their very extensive distribution around the globe, the monitoring of those resources represents a major component of environmental assessments. It is now well recognised that remote sensing is a prime source of relevant information in this respect.
Because of its characteristics the VEGETATION instrument occupies a special position in the panoply of remote sensing approaches for forest monitoring. It is obvious that with a resolution of 1000 meters, the instrument cannot address issues related to local management. What can be made visible in a one kilometre "pixel" is related to the type of vegetation and its current conditions more than to the characteristics of the forest stands themselves. As an example the project "VEGETATION data in the monitoring of disturbances patches and post-disturbance succession in natural vegetation in southern Africa" aims to introduce VEGETATION data in a vegetation monitoring and mapping methodology. In this case, the data is expected to help in obtaining information about the short term history of individual regions in terms of major fires, floods and other disturbances.
Nevertheless, because of its geographical breath and frequency of coverage that instrument will support observations which are not efficiently done using high resolution satellite data sets. One must see the future VEGETATION programme as an additional element of information on the distribution and state of the world forests at regional to continental and global scales.
Currently, the scientific community and, among others, the TREES project of the Joint Research Centre has obtained appreciable results in global forest monitoring using the existing low resolution AVHRR 1km data. The use of such data is twofold: first, identification and mapping of tropical forest cover types and, second, assessment of forest seasonality. The first activity is based upon the analysis of individual satellite images found in the collection of daily passes collected over the whole tropical belt. This has required the running of more than seven receiving stations around the world. In order to minimise the impact of clouds on the data set a near constant observation was maintained (daily acquisitions from the end of the rainy season to the end of the dry season). The African, Asian and Latin American continents were thus covered and the result consists in a world forest map obtained with a single sensor and a uniform forest cover classification. An example of the results obtained using those AVHRR 1km data is given for Rio Branco (Brazil). At such low resolution levels, the results need to be rigorously validated. In the present case this was done by "correcting" the AVHRR analysis using satellites with a higher resolution (in this case Landsat Thematic Mapper- 30 meters and SPOT HRVIR). It is worth pointing here at the fact that the combined use of low and high resolution satellite data is very much at the core of global vegetation monitoring. The perspectives offered by the simultaneous acquisition of VEGETATION - SPOT HRVIR data acquisition will allow for a better operation of this validation procedure.
Vegetation seasonality has gained a renewed interest in the framework of global vegetation dynamics studies. Seasonality is, indeed, a characteristic which regulates many processes such as those related to primary productivity, water and gas exchanges. Being directly linked to climate conditions, seasonality is a topic of prime relevance in the assessment of potential climate impacts; these impacts could be theoretically identified as anomalies or unexpected trends in a seasonal signal produced by a vegetation canopy. Furthermore, radiometric changes in the canopy associated with the succession of the seasons is a typological characteristic of the vegetation itself (see for ex. classical references to seasonal, evergreen forests etc.). Thus the availability of well calibrated long-term time series of canopy measurements is of much value for classifying vegetation. The VEGETATION instrument will significantly improve upon the current performance of the AVHRR to obtain such consistent data sets.
Therefore, it is foreseen that the data which will be regularly provided by the VEGETATION instrument will be directly assimilated into the current and future observation systems dedicated to the monitoring of the forests of the world.